# Task Metadata Audit > Use when task items or Kanban links need a read-only check for readiness, dependencies, issues, acceptance, or board consistency. - **Type:** Skill - **Install:** `agentstack add skill-choral-io-choral-skills-task-metadata-audit` - **Verified:** Yes — security-reviewed for prompt injection and unsafe behavior - **Seller:** [choral-io](https://agentstack.voostack.com/s/choral-io) - **Installs:** 0 - **Category:** [AI & ML](https://agentstack.voostack.com/c/ai-and-ml) - **Latest version:** 0.1.0 - **License:** MIT - **Upstream author:** [choral-io](https://github.com/choral-io) - **Source:** https://github.com/choral-io/choral-skills/tree/main/skills/task-metadata-audit - **Website:** https://github.com/choral-io/choral-skills ## Install ```sh agentstack add skill-choral-io-choral-skills-task-metadata-audit ``` Requires the [AgentStack CLI](https://agentstack.voostack.com/docs/cli). Works with Claude Code, Cursor, and any MCP-compatible agent. ## About # Task Metadata Audit ## Runtime Context Before acting, resolve `` using the runtime bootstrap rules, then read `/.workflow/runtime.md` and `/.workflow/manifest.yml`; do not assume non-default workflow paths or default ids. Use this skill to inspect task metadata and report issues. This skill is read-only. ## Workflow 1. Read `/.workflow/schemas/common.md`. 2. Read `/.workflow/schemas/tasks.md`. 3. Read `/.workflow/rules/delivery.md`. 4. Read `/planning/KANBAN.md`. 5. Scan `/tasks/*.md`, excluding localized files. 6. Parse frontmatter and task body sections. 7. Check task relationships against existing task ids and Kanban card ids. 8. Report issues and dry-run fixes without editing files. ## Checks - Missing or invalid `type`, `priority`, `value`, `effort`, `readiness`, or `module`. - Missing or invalid `severity` only when `type` is `issue`, `bug`, or `defect` and impact is needed for triage. - Missing, empty, or unresolved `owners` in task items that are `readiness: ready`, linked from a Kanban card, scheduled, assigned, or actively maintained. - `owners`, `assignees`, or `reviewers` values that reference member or group documents without wikilinks, or tool-written values that should be path-qualified. - Group values in `assignees`; report these as team-pool assignments, not direct assignment to the current member. - Missing `assignees` only when the task appears scheduled or actively assigned. - Missing `reviewers` only when review responsibility is expected. - `blocked_by` or `related_to` entries that point to missing or ambiguous task references. - Missing `## Sources` section or sources that point only to local workspace notes or localized files. - Tasks with `readiness: ready` but missing acceptance criteria. - Issue, bug, or defect tasks with `readiness: ready` but missing `Problem`, `Impact`, or `Triage` context. - Reproducible bug or defect tasks with `readiness: ready` but missing `Reproduction`, `Expected`, or `Actual`. - Tasks with `readiness: ready` whose task item or required source materials are uncommitted. - Tasks with `readiness: ready` whose task item or required source materials are not pushed to the default remote when one exists. - Tasks in `Ready` with unresolved `blocked_by` entries. - Tasks in `Ready` with `readiness` other than `ready`. - Tasks in `Blocked` without `readiness: blocked`, unresolved `blocked_by`, or an explicit blocker note. - Tasks in `Backlog` with `readiness: blocked`; report as allowed planned dependency state, not an error. - Tasks outside `Backlog` or `Blocked` with unresolved `blocked_by` entries. - Downstream tasks whose `blocked_by` references are all resolved but still have `readiness: blocked` or remain in `Blocked`. - Tasks in `Backlog` that appear ready but have not been promoted to `Ready`. - Localized task files used as planning inputs. ## Output Use the standard task audit output shape from `references/report.md`: `Summary`, `Findings`, `Dry-run fixes`, `Requires judgment`, `Safe for next-task-selection`, and `Sources`. ## Guardrails - Do not edit task items. - Do not move Kanban cards. - Do not infer sensitive information. - Do not treat `assignees` as Kanban status. - Treat source stability as a readiness audit and Ready-column maintenance check. Do not require `next-task-selection` to run the full audit for every candidate unless the user asks for strict verification or automatic start. - Require maintainer approval before any suggested fix is applied by another skill. ## References - For audit output examples, read `references/report.md`. ## Source & license This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code. - **Author:** [choral-io](https://github.com/choral-io) - **Source:** [choral-io/choral-skills](https://github.com/choral-io/choral-skills) - **License:** MIT - **Homepage:** https://github.com/choral-io/choral-skills Install and usage instructions live in the source repository linked above. ## Pricing - **Free** — Free ## Security capabilities Automated source analysis of v0.1.0 — what this tool can access: - **Network access:** no - **Filesystem access:** no - **Shell / process execution:** no - **Environment & secrets:** no - **Dynamic code execution:** no *"Yes" means the capability is present in the source — more access means more to trust, not that it is unsafe.* ## Versions - **0.1.0** — security scan: passed — Imported from the upstream source. ## Links - Listing page: https://agentstack.voostack.com/l/skill-choral-io-choral-skills-task-metadata-audit - Seller: https://agentstack.voostack.com/s/choral-io - Browse the marketplace: https://agentstack.voostack.com/browse --- Listed on AgentStack — the marketplace for AI agent skills and MCP servers. Every listing is security-reviewed. Creators keep 70%.